Gene of Klebsiella Oxytoca

Biosci. Biotechnol. Biochem., 72 (1), 116–123, 2008

Identification of the 4-Hydroxycinnamate Decarboxylase (PAD) Gene of Klebsiella oxytoca

y Hirofumi UCHIYAMA, Yasuyuki HASHIDOKO, Yuhki KURIYAMA, and Satoshi TAHARA

Division of Applied Bioscience, Research Faculty of Agriculture, Hokkaido University, Kita-ku, Sapporo 060-8589, Japan

Received August 3, 2007; Accepted October 15, 2007; Online Publication, January 7, 2008 [doi:10.1271/bbb.70496]

A 7.1-kbp DNA fragment isolated from a wild strain also a synonym of phenylacrylic acid decarboxylase, of Klebsiella oxytoca was sequenced, leading to the PAD) to utilize the decarboxylative product, 4-hydroxy- identification of 10 open-reading frames (ORFs), in- lated styrene, as an antifungal agent.3) In 1994, Hashi- cluding a 504-bp Pad gene. The Pad gene of the Gram- doko et al. successfully cloned the Pad gene from a negative bacterium was subsequently expressed in genomic library prepared from wild-type K. oxytoca.4) A Escherichia coli as a chimeric Pad. The deduced amino 9.6-kbp genomic fragment carrying the Pad gene was acid (AA) sequence of the Pad gene from wild-type obtained using Cladosporium herbarum AHU 9262 as a K. oxytoca showed approximately 50% homology to bioindicator which was highly sensitive to 4-hydroxy- those of other bacterial PADs from Gram-positive styrene produced from 4-hydroxycinnamic acid via PAD bacilli plus a coccus. These data and a genomic library catalytic decarboxylation. Transformed Escherichia search of some -proteobacteria, including E. coli and coli JM 109 harboring the cloned Pad gene region Vibrio sp., indicated that PAD of K. oxytoca is a member acquired the ability to decarboxylate (E)-4-hydroxycin- of the bacterial PAD family characteristic of Gram- namate into 4-hydroxystyrene. Digestion with SacI and negative bacteria. Using Pad-specific PCR primers EcoRI cleaved this 9.6-kbp DNA region into three designed from the Gram-negative bacterial Pad of fragments: HindIII-EcoRI (2.4 kbp, fragment C), EcoRI- K. oxytoca, Pad genes of two further strains of K. oxy- SacI (4.7 kbp, fragment B), and SacI-HindIII (2.5 kbp, toca, another wild isolate and JCM 1665 and two PAD- fragment A). These three fragments were independently positive Enterobacter spp. were successfully amplified cloned into the pUC19 vector. The expression of Pad in for specific Pad detection. each fragment in transformed E. coli was unsuccessful. Only bacterial cells transformed with recombinant DNA Key words: Klebsiella oxytoca; bacterial Pad gene; PCR containing fragment CB (7.1 kbp, HindIII-SacI possess- detection of Pad; Gram-negative bacteria; ing an EcoRI site inside the DNA sequence) gained the 4-hydroxycinnamate decarboxylase gene ability to decarboxylate 4-hydroxycinnamic acid (un- published data), but the DNA sequence of the Pad gene Some plant families, such as Convolvulaceae, Aster- on fragment CB has not yet been characterized. iaceae, and Umbelliferae, are representative chlorogenic Bacterial Pad gene studies have been done in food acid-accumulating plant families that contain high chemistry, due to serious issues of off-flavor during the concentrations of hydroxycinnamoyl quinic acids in fermentation of foods and liquor. In 1995, bacterial Pad their leaves and/or roots. Many of these plant poly- of Gram-positive Bacillus pumilis was first characterized phenols are now recognized as self-defensive substan- by Zago et al., and a full-length amino acid (AA) ces, but the roles of these plant products in interactions sequence of PAD was deduced.5) Since this identifica- between polyphenol-rich plants and host-associating tion of Pad, additional Pad genes have been charac- microorganisms are not well understood.1,2) In 1993, terized in more than 20 different microorganisms, but Hashidoko et al. investigated phylloplane bacteria from almost all of these microorganisms are Gram-positive the chlorogenic acid-accumulating plant Polymnia son- bacilli plus a coccus.6–8) In contrast, among Gram- chifolia, a member of the Asteriaceae family. A phyllo- negative bacteria, only Pseudomonas fluorescens9) and plane bacterium was tentatively identified as Klebsiella K. oxytoca10) have been reported to possess PAD oxytoca by morphological and biochemical character- activity in crude protein from bacterial cells. Bacterial ization. This organism was found actively to decarbox- Pad-like ORFs have been found in genomic DNA from ylate 4-hydroxylated cinnamic acids by a substrate- two Gram-negative proteobacteria, Vibrio cholerae and inducible 4-hydroxycinnamate decarboxylase (4-HCD; Erwinia carotovora,11,12) but it is uncertain whether the

y To whom correspondence should be addressed. Fax: +81-11-706-4182; E-mail: [email protected] Identification of Pad Gene from Klebsiella oxytoca 117 encoded gene products possess PAD functionality in Characterization of ORF regions in fragment CB. The vivo. Hence both the enzymatic and the bacterio- target DNA (fragment CB), possessing 7,102 bp between physiological traits of PAD have yet to be examined the HindIII and SacI sites, did not contain any regions in Gram-negative bacteria. In fact, the physiological and that exhibited high homology to Pad of Bacillus spp. ecological roles of PAD-producing bacteria and PAD Hence all of the ORF regions on this DNA fragment itself in the terrestrial ecosystem are unclear, because were searched by means of an ORF Finder tool provided K. oxytoca immediately hydrolyzed chlorogenic acid to by NCBI (http://www.ddbj.nig.ac.jp/search/blast-j.html). yield 4-hydroxylated styrenes in chlorogenic acid-con- ORF regions were searched from both plus and minus taining potato-dextrose medium.13) In this study, we strand directions with putative amino acid translations aimed to determine the DNA sequence of Pad in from three different frames (+1, +2, +3, 1, 2, and fragment CB originating in epiphytic K. oxytoca, and to 3). A total of 20 ORF candidates of 250 bp or greater describe the PADs of other strains of K. oxytoca and two were identified on the target DNA. The deduced amino further wild isolates of Enterobacter spp. acid (AA) sequences of the putative proteins encoded on each ORF were further searched on the BLASTX Materials and Methods database of DDBJ (http://www.ddbj.nig.ac.jp/search/ blast-j.html). General. The 7.1-kb DNA fragment (fragment CB) carrying the gene for PAD was identical to one Design of adapter primers for Pad gene insertion into previously cloned from chromosomal DNA of a wild- pGEX 4T-3. In order to sub-clone target ORF regions into type K. oxytoca bacterial strain. This DNA fragment the pGEX vector, adapter primers were designed on the possessed HindIII and SacI sites at the 50 and 30 ends SIGMA Genosys website (http://www.genosys.jp/adt/ respectively, and it was cloned into pUC19 vector SGJ NN 3rd.html), and predictions for Tm, hairpin, and plasmid. For PCR, HotStarTaq (Qiagen, Germantown, dimer structures were obtained. The designed adapter MD) was used at 97 C for 15 min for preliminary primers were: PAD-APF (forward, 50-CGGAATTC- heating. The reaction conditions are described in detail TATGAGCACATTCGACAAACA-30, 29 mer, Tm ¼ for each experiment below. DNA and amino acid (AA) 73 C), and PAD-APR (reverse, 50-CGCTCGAGGAT- sequence homology searches were performed on the TACAGGTTGGCAGGAA-30, 27 mer, Tm ¼ 77 C). website of NCBI (National Center for Biotechnology These primers possessed EcoRI and XhoI restriction Information, USA) and DDBJ (DNA Data Bank of sites at the 50 and 30 ends respectively. The specificity of Japan, Mishima, Japan). the primers was confirmed by PCR using fragment CB as the source of template DNA. PCR conditions were as Primer design for direct sequencing. The primers follows for the amplification of the ORF-7 regions: 30 used in this study are shown in Table 1. Based upon the cycles of denaturation at 94 C for 1 min, annealing at sequence determination obtained from the initial round 50 C for 1 min, and extension at 72 C for 1 min. of sequencing, we designed customized forward/reverse The resulting PCR product (523 bp) was double- internal primers. The DNA sequence of fragment CB digested with EcoRI and XhoI, and the enzymes were was registered in the DDBJ DNA data bank under denatured by heating. The digested PCR product was accession no. AB330293. then ligated into pGEX 4T-3 expression vector (Amer- sham Pharmacia Biotech, Uppsala, Sweden) at the EcoRI-XhoI cloning site for in-frame fusion to the Table 1. Primers Used in Sequencing of Fragment CB internal GST gene in pGEX 4T-3. A set of adapter primers (1, AP3, and AP4) was designed for Pad 50-GTTCCACGGTCTCTTCC-30 (160–176) 50-GATGTTGATGATCTGCC-30 (395–411) expression. The ligation reaction was set at a molar ratio 50-GGATAATCTTCAGGTTCTC-30 (837–856) of the vector plasmid:PCR product = 1:10. E. coli cells 50-GTTTCGCTTTCACCGT-30 (1,707–1,722) were transformed with the recombinant plasmid to 50-ATGTCATCGCACAGCA-30 (1,815–1,830) 0 0 facilitate the production of recombinant, PAD GST- 5 -CATCCCGACGAACTGG-3 (2,208–2,221) fusion protein. Transformed cells were assayed for PAD 50-AGAGCGATGACGCTGCCGAA-30 (2,336–2,355) 50-GGATCTCAACCG-30 (2,474–2,486) activity with a decarboxylation test in the cultured 10) 50-GTTGATCATCTTCGAGG-30 (2,635–2,651) medium using (E)-caffeic acid as the substrate. 50-GCTTCGGTGAGCAT-30 (2,888–2,901) 50-GAGTCATCGCTGCT-30 (2,909–2,922) Confirmation of PAD activity in transformed 0 0 E. coli 5 -AGCTCGCGCAGATAG-3 (3,399–3,413) with recombinant pGEX 4T-3 plasmid. E. coli JM109 50-GGATAATTGCTTTGGTTC-30 (3,832–3,849) 50-CGTGAAACAGTTTATCG-30 (4,843–4,859) competent cells (Nippon Gene, Toyama, Japan) were 50-GTCGGTGAGCTATAATCTC-30 (5,412–5,430) transformed with the recombinant pGEX 4T-3PAD 50-ATGCGGTGCTGTATCAG-30 (5,974–5,990) plasmid carrying the Pad gene. Hundreds of colonies 50-CGCCGAATGATATCAC-30 (6,418–6,433) 0 0 emerged on a selection plate containing 50 mg/liter of 5 -GCTCCGTCGCAGTTCAAAGA-3 (6,840–6,859) ampicillin. Among these, 96 colonies (B1-B96) were 50-ATGGGTACTGAAGG-30 (6,905–6,918) taken randomly from the selection plates and subjected 118 H. UCHIYAMA et al. Table 2. ORFs on Fragment CB

no Frame Position Size Protein 1 1 2083–98 1986 NAD-dependent epimerase/dehydratase, formyl transferase 2 1 4315–3065 1251 UDP-4-amino-4-deoxy-L-arabinose oxoglutarate aminotransferase 3 þ2 152–1264 1113 Not hit 4 2 3117–2080 1038 Glycosyl transferase 5 þ2 6248–7102 855 Bacterial extracellular solute-binding protein 6 þ1 2530–3099 570 Carboxyl esterase (Burkholderia sp.) 7 1 5107–4604 504 Bacterial phenolic acid decarboxylase (Pad) 8 þ2 3221–3676 456 Not hit 9 þ2 5645–6067 423 Transcriptional regulator LysR family 10 2 1665–1251 405 Not hit 11 þ2 2171–2569 399 Not hit 12 þ3 5208–5588 381 Transcriptional regulator LysR family 13 1 6658–6278 381 GTP binding protein (only for domain near N-terminal) 14 þ2 4769–5131 363 Not hit 15 þ2 1772–2101 330 Not hit 16 þ3 1545–1835 291 Not hit 17 1 7102–6815 288 Ni-binding periplasmic protein 18 þ1 79–363 285 Not hit 19 þ3 771–1055 285 Not hit 20 þ3 390–656 267 Not hit

All of the ORFs greater than 250 bp are listed. ORF-7 encoded a putative protein that showed high sequence homology to bacterial PADs of Gram-positive bacteria. to an insert check. Positive clones were then tested for K. oxytoca 60E, two Enterobacter spp. (70S and 136A), decarboxylation activity. In this assay, (E)-caffeic acid and two Burkholderia spp. (67A and 67R). Among the was added as the substrate of PAD to 12-h-cultured tested Gram-negative bacteria, K. oxytoca JCM 1665, medium to obtain a final concentration of 1 mM. After and K. oxytoca 60E possessed a substrate-dependent, further incubation for 30 min, the medium was extracted inducible type of PADs. On the other hand, Enter- with EtOAc, concentrated, and developed on TLC obacter spp. 70S and 136A possessed a constitutive type (Merck Kieselgel F254 Art 5715, 0.25 mm thick, Merck, of PAD, the latter of which was relatively weak in Darmstadt, Germany) in CHCl3-MeOH = 9:1 to check decarboxylation activity. Burkholderia spp. showed for the presence of 3,4-dihydroxystyrene. no PAD activity even in the presence of substrates, but they showed remarkable gallate decarboxylation activity Recovery of the Pad gene fragment from the PAD equal to Rhanella aquatilis.15) K. oxytoca 60E, Enter- positive clone. The PAD assay-positive clones were obacter sp. 70S, and Enterobacter sp. 136A were cultured in LB-ampicillin medium for 24 h at 37 C. The originally isolated from the roots of Chenopodium recombinant plasmid copied was recovered from the album, Rudbeckia laciniata, and C. album respectively. cultured E. coli cells by the lysozyme-alkaline method. Two Burkholderia spp., 67A and 67R, were from a The resulting plasmid was subjected to double-digestion Polygonum sp. with EcoRI/XhoI to obtain about 500 bp of inserted DNA. This fragment was subsequently sequenced (ABI Results PRISM 310 Genetic Analyzer, Applied Biosystems, Foster City, CA) in order to confirm the amplified DNA Identification of bacterial Pad from Klebsiella oxy- to be Pad from ORF-7. toca Full DNA sequencing of fragment CB (7.1 kb, Design of degenerate PCR primers for reliable Pad HindIII-SacI fragment from wild-type K. oxytoca) re- gene detection by PCR. The characterized Pad gene of sulted in the identification 20 open reading frames K. oxytoca possessed some common motifs with those (ORFs, ORF-1-20) that were greater than 250 bp in size. of Bacillus spp.,5,7) Lactobacillus spp.,6) and Pediococ- Among these, 11 ORFs encoded proteins/enzymes of cus pentosaceus.8) Bacterial Pad gene-detecting primers known function (Table 2). Partial N-terminal amino acid were designed from common motifs common to these (AA) sequences of ferulic acid decarboxylases purified organisms. The degenerate primers were then tested on from L. plantarum and Pseudomonas fluorescens16,17) Gram-negative bacteria isolated from the rhizosphere of exhibited 50% homology to the AA sequence encoded in wild plants collected throughout the Hokkaido Univer- ORF-7 (5,107–4,604, on the reverse chain). The putative sity Campus.14) Together with K. oxytoca JCM 1665, amino acid sequence encoded in ORF-7 was searched on used as a reference bacterium,13) these five newly InterProScan (http://www.ebi.ac.uk/InterProScan/).18) isolated bacteria were tested for PCR detection of The 504 bases of ORF-7 were characterized as the bacterial Pads. These were tentatively identified as Pad gene in view of nearly 50% agreement of the coding Identification of Pad Gene from Klebsiella oxytoca 119

Fig. 1. Differences in PAD Characters between Gram-Positive Bacilli and Gram-Negative Rods. A, The AA motifs indicated by colored backgrounds are preserved AA sequences among bacterial PADs. Using three motifs out of four, degenerate primers were designed for bacterial Pad-detecting PCR. KO, Klebsiella oxytoca; PP, Pediococcus pentosaceus; LP, Lactobacillus plantarum; BS, Bacillus subtilis; BP, Bacillus pumilus. , identical amino acids. B, Oligonucleotide sequences of degenerate primers for bacterial PADs to cover Gram-negative bacteria. C, Bacterial PAD-detecting PCR assay for some test bacteria. Left, PCR products with PCD- F1/PCDR1; right, PCR products with PCD-F1/PCD-R2 primers. Electrophoresis was performed in a 4% stacking and 13% separating polyacrylamide gel. M, marker is a 100-bp ladder marker (100–1,000 and 1,500); 1, Klebsiella oxytoca 60E; 2, K. oxytoca JCM 1665; 3, Burkholderia sp. 67A; 4, Burkholderia sp. 67R; 5, Enterobacter sp. 70S; and 6, Enterobacter sp. 136A. Both PCR detections, with PAD-F1/ PAD-R1 and PAD-F1/PAD-R2, were performed under the following PCR conditions: 35 cycles of denaturation at 94 C for 1 min, annealing at 50 C for 1 min, and extension at 72 C for 1 min. The volume of each primer solution (50 mM) added to the reaction solution (50 ml) was 1.5 ml. The resulting PCR products were run on a 13% polyacrylamide gel by slab gel electrophoresis along with a 100-bp ladder marker.

AA sequences with PADs reported from Gram-positive positive on caffeic acid decarboxylation assay (Fig. 2). bacilli and a coccus. Furthermore, some conserved The transformed E. coli JM109/pGEX-4T-3-PAD clone motifs common to PAD family proteins were identified B43 was more active than pUC19-CB-transformed in the PAD gene (Fig. 1A).5–8) E. coli JM109. B35 and B43, selected as PAD-positive After full sequencing of fragment CB, it was evident clones, were subsequently confirmed by PCR to carry that ORF-7 was located in reverse orientation to the lac the Pad gene in pGEX-4T-3-PAD plasmid. Expression promoter in the pUC19 plasmid. In addition, the sole of the Pad gene from Gram-positive bacteria in E. coli EcoRI site was located at positions, 2,449–2,454, a has been reported,19) but this is probably the first region that was a long distance from ORF-7. Since example of gene expression of a Gram-negative bacte- neither fragment C nor B showed PAD activity on each rial Pad in E. coli. recombinant plasmid when fragment CB was digested with EcoRI, we expected that the Pad gene would Degenerate PCR for detection of the Pad gene using contain the EcoRI site, but this EcoRI site was not newly designed Pad-specific primers related to Pad gene expression. It is perplexing that the The amino acid sequence of the characterized PAD of pUC plasmid carrying fragment C did not allow for the wild-type K. oxytoca indicated that it is a decarboxylase expression of PAD in transformed E. coli. On the other of the bacterial PAD family. Sequence analysis also hand, it is reasonable in view of the reverse direction of confirmed the presence of some motifs common to those the Pad gene against the lac promoter that IPTG failed of Bacillus spp. and Lactobacillus plantarum.6) For the to activate Pad expression in cells transformed with the characterization of PADs possessing motifs similar to recombinant pUC19-CB-carrying Pad gene. those of Gram-positive bacilli, detection of Pad in wild- Since we failed to observe induction in the pUC19 type K. oxytoca and K. oxytoca JCM 1665 by DNA system, we inserted the ORF-7 region into the pGEX- homology searches failed. Furthermore, PCR detection 4T-3 expression vector to obtain 10 independent with degenerate primers designed from Gram-positive colonies. The two clones (B35 and B43) tested were bacilli were unsuccessful (data not shown) until the ORF 120 H. UCHIYAMA et al.

CHCl3-MeOH 9:1

indole

HO

HO

HO COOH 3,4-dihydroxystyrene

HO

(E )-caffeic acid CA 1234567

Fig. 2. PAD Activity of Transformed E. coli with the Pad Gene. E. coli JM 109 that possessed ORF-7 inserted into pGEX-4T-3 expression vector plasmid (pGEX-4T-3-PAD). CA, caffeic acid only (5 mM in 10 ml of water); 1, LB broth medium only (10 ml); 2, LB medium + caffeic acid; 3, E. coli JM109/pUC19 + IPTG and caffeic acid; 4, E. coli JM109/pGEX-4T-3 + IPTG and caffeic acid; 5, E. coli JM109/pUC19-CB + IPTG and caffeic acid (positive control); 6, E. coli JM109/pGEX-4T-3-PAD B40 + IPTG and caffeic acid (negative control); 7, E. coli JM109/pGEX-4T-3-PAD B43 + IPTG and caffeic acid (positive). In all portions, IPTG was added to the cultures from the initiation of incubation. After overnight culture, 5 mM of caffeic acid was added as the decarboxylation substrate, and these transformed E. coli were further incubated for 30 min. After adjusting to pH 4.0 with 1 M HCl, 2 ml of EtOAc was added to the medium, and the test tube was vortexed for 1 min. The resulting EtOAc layer was then analyzed by TLC developed in CHCl3-MeOH 9:1. search in fragment CB. Hence, we designed new bacterial Pad gene detection. Using the fragment CB as degenerate primers from common motifs of AA se- a positive control, the combination of PAD-1F/PAD-1R quences that are encoded on bacterial Pad genes and and PAD-1F/PAD-2R produced amplified PCR prod- Pad-like pseudo-genes. ucts of 172 bp and 242 bp respectively. Based on comparisons among AA sequences of some Consequently, two K. oxytoca and two Enterobacter bacterial PADs, including that of phytoepiphytic K. oxy- spp. tested by PCR with the combination PAD-1F/PAD- toca and ones obtained in a database search, YTY- 1R produced products of reasonable fragment sizes DNGW (19–25 AA, for forward primer), WTEPTGT (> 200 bp), but the PCR products of two Burkholderia (70–76 AA, for 1st reverse), and IFFPRWV (93–99 AA, spp. were 300 bp in size (Fig. 1C, left). On the other for 2nd reverse) were selected as the target motifs in the hand, while four -proteobacteria of K. oxytoca (JCM wild-type K. oxytoca PAD to design Pad-specific 1665 and 60E) and Enterobacter spp. (70S and 136A) primers (Fig. 1A). In this primer design, YTYDNQW tested positive on PCR assay with PAD-1F/PAD-2R, and IYFPRWI of Vibrio cholerae (AE004296) PAD, two Burkholderia spp. were negative (Fig. 1C, right). located at 19–25 AA and 93–99 AA respectively,12) and The PCR products obtained by PAD-1F/PAD-2R YTYANGW and IFFPQWI of Mycobacterium avium primers were sequenced and compared for homology subsp. paratuberculosis (AE017229) at 28–34 AA and to their respective PADs on the amino acid level 102–108 AA respectively20) were not included. These (Fig. 3). The PCR products of K. oxytoca JCM 1665 and regions were excluded because these bacteria were 60E, both of which possessed an inducible type of active assumed to be neither soil nor plant-associating bacteria. PADs, and Enterobacter sp. 70S with a constitutive From the preserved AA motifs of the target PAD, a PAD exhibited high homology to the partial AA se- degenerate forward (PAD-F1, 50-TAYACNTAYGANA- quence of bacterial PAD that was coded on the fragment AYGGNTGG-30) and two reverse (PAD-R1, 50-GTN- CB from the phylloplane K. oxytoca. In contrast, Enter- CCNGTNGGYTCNGTCCA-30 and PAD-R2, 50-ACC- obacter sp. 136A, which possesses the constitutive type CANYKNGGRAARAADAT-30) primers encoding the of PAD, exhibited low AA sequence homology to that of AA motifs were designed (Fig. 1B). Many of the PAD- the inducible-type of Gram-negative bacteria. positive bacteria isolated from the rhizoplanes of several Altogether, our degenerate primers, designed spe- plants in our systematic screening in the field showed cifically for bacterial PAD gene detection, were appli- decarboxylation activities almost equivalent to a refer- cable to -proteobacteria (genera Klebsiella and Enter- ence strain, K. oxytoca JCM 1665. Five wild strains of obacter), but it is necessary to determine whether these Gram-negative eubacteria, isolated from the root sur- primers are specific only to bacterial Pad of -pro- faces of various host plants, were subjected to PCR teobacteria (consisting mainly of inducible types of the assay using these degenerate primers for Gram-negative PAD group), or whether they have broad specificity to Identification of Pad Gene from Klebsiella oxytoca 121

Fig. 3. Encoding Partial AA Sequences of the PCR Products from Bacterial Pad. AA sequences shown by a light-gray background were common AA with substrate-inducible PADs of K. oxytoca, a Gram-positive bacterium. AA sequences alternative to those of K. oxytoca are shown by a dark-gray background, in which AA sequence variations were confirmed among Gram-positive bacteria and constitutive PADs or pseudo-gene-like Pad-possessing Gram-negative bacteria.

several bacterial genera and subclasses. In the PCR capable of inducing PADs under exposure to substrates, assay and subsequent DNA sequence determination, a or why it is necessary to accumulate such decarbox- constitutive type of bacterial PAD of Enterobacter sp. ylative products in their habitat. Probably, de novo 136A was found to possess a unique AA sequence synthesis in PAD production is a trade-off between region. In other words, substrate-inducible PADs of - detoxification and energy consumption. proteobacteria, at least, possess common motifs (the A In our current research, Pad characterization from and B regions as shown in Fig. 3). Because many of the K. oxytoca, a Gram-negative phytoepiphyte (phyllo- constitutive types of PAD-possessing Gram-negative plane bacterium) of -proteobacteria was done first, bacteria have been isolated (data not shown), it is along with sequence determination of the upstream and necessary to determine whether constitutive PAD is downstream regions of the Pad gene. This is significant, separated into sub-families of PAD among Gram- because it will be possible to focus on important DNA negative bacteria. In our bacterial collection, they regions associated with the regulation mechanism of showed a diverse and wide range of variations of substrate-dependent, inducible Pad gene expression in enzymatic behaviors of PADs.13,21) A different group of K. oxytoca. Since substrate-specificity and substrate- PADs has been reported in ferulate decarboxylases of dependent induction of PAD has been well studied in yeast (named yeast PAD1). Since this group of Pad K. oxytoca JCM 1665, genetic information from frag- genes do not exhibit any AA sequence homology to the ment BC is probably applicable to JCM 1665. In fact, it bacterial PADs,22,23) it is speculated that the origins of has been found that an indigested substrate analog, 6- the yeast and the bacterial Pad genes are completely hydroxynaphthoic acid, acts in K. oxytoca JCM 1665 different. In fact, a function-unknown UbiX-like decar- cells as a powerful PAD inducer, maintaining its boxylase (PAD1) of Escherichia coli O157:H7 possess- induction activity longer than (E)-caffeic acid or other es an AA sequence that is highly homologous to those of natural substrates. Thus, both chemical and genetic tools yeast PAD1,24) but E. coli does not show any PAD-like for further biochemical study became available. Bar- activity that is either constitutive or inducible. thelmebs et al. have reported as a pioneer work that PadA gene expression in Pediococcus pentosaceus is Discussion regulated by bicistronic transcription of a PadR gene located downstream of PadA. In this bacterium, the PADs produced by Gram-negative bacteria are PadR protein acts as a transcriptional repressor of the grouped into two types: inducible PADs and constitutive PadA operon upstream of the PadA gene. It is important PADs.21) In our investigations, bacteria possessing to note that similar repressor regions were not found constitutive PADs tended to be weakly active, and upstream or downstream of ORF-7 in fragment BC.8) conversively, bacteria that contain substrate-inducible The Pad gene-specific primers that were developed in PADs were highly active on decarboxylation assay (data this study should serve as a useful Pad-detection tool not shown), but it is not clear why some bacteria are also for study of the environmental roles of the PAD- 122 H. UCHIYAMA et al. producing bacteria. Pad gene distribution and expres- A., Gwinn, M. L., Dodson, R. J., Haft, D. H., Hickey, sion in phenolic compound-accumulated habitats of E. K., Peterson, J. D., Umayam, L., Gill, S. R., Nelson, phytoepiphytic bacteria indicate the importance of the K. E., Read, T. D., Tettelin, H., Richardson, D., PAD reaction in the terrestrial ecosystem. Ermolaeva, M. D., Vamathevan, J., Bass, S., Qin, H., Dragoi, I., Sellers, P., McDonald, L., Utterback, T., Fleishmann, R. D., Nierman, W. C., White, O., Salzberg, Acknowledgments S. L., Smith, H. O., Colwell, R. R., Mekalanos, J. J., Venter, J. C., and Fraser, C. M., NA sequence of both We thank Mr. Albert Asante of our laboratory for chromosomes of the cholera pathogen Vibrio cholerae. providing some of his rhizobacterial collections. This Nature, 406, 477–483 (2000). research was supported by a Grant-in-Aid for Scientific 12) Bell, K. S., Sebaihia, M., Pritchard, L., Holden, M. T. G., Research from JSPS (The Japan Society for the Hyman, L. J., Holeva, M. C., Thomson, N. R., Bentley, Promotion of Science) (no. 16208032, to Y.H.). S. D. L., Churcher, J. C., Mungall, K., Atkin, R., Bason, N., Brooks, K., Chillingworth, T., Clark, K., Doggett, J., References Fraser, A., Hance, Z., Hauser, H., Jagels, K., Moule, S., Norbertczak, H., Ormond, D., Price, C., Quail, M. A., 1) Kosuge, T., The role of phenolics in host response to Sanders, M., Walker, D., Whitehead, S., Salmond, G. P. infection. Annu. Rev. Phytopathol., 7, 195–222 (1969). C., Birch, P. R. J., Parkhill, J., and Toth, I. K., Genome 2) Friend, J., Phenolic substances and plant disease. Annu. sequence of the enterobacterial phytopathogen Erwinia Proc. Phytochem. Soc., 25, 367–372 (1985). carotovora subsp. atroseptica and characterization of 3) Hashidoko, Y., Urashima, M., Yoshida, T., and virulence factors. Proc. Nat. Acad. Sci. USA, 101, Mizutani, J., Decarboxylative conversion of hydroxy- 11105–11110 (2004). cinnamic acids by Klebsiella oxytoca and Erwinia 13) Hashidoko, Y., and Tahara, S., Stereochemically specific uredovora, epiphytic bacteria of Polymnia sonchifolia proton transfer in decarboxylation of 4-hydroxycinnamic leaf, possibly associated with formation of microflora on acids by 4-hydroxycinnamate decarboxylase from Kleb- the damaged leaves. Biosci. Biotechnol. Biochem., 57, siella oxytoca. Arch. Biochem. Biophys., 359, 225–230 215–219 (1993). (1998). 4) Hashidoko, Y., Urashima, M., and Mizutani, J., Cloning 14) Asante, A., Tahara, S., and Hashidoko, Y., Screening of of a DNA fragment carrying the 4-hydroxycinnamate rhizobacteria possessing phenolic acid-decarboxylation decarboxylase (PofK) gene from Klebsiella oxytoca, and abilities from several plant families. J. Res. Fac. Agric. its constitutive expression in Escherichia coli JM109 Hokkaido Univ., accepted (2008). cells. Biosci. Biotechnol. Biochem., 58, 217–218 (1994). 15) Hashidoko, Y., Itoh, E., Yokota, K., Yoshida, T., and 5) Zago, A., Degrassi, G., and Bruschi, C. V., Cloning, Tahara, S., Characterization of five phyllosphere bacteria sequencing, and expression in Escherichia coli of the isolated from Rosa rugosa leaves, and their phenotypic Bacillus pumilus gene for ferulic acid decarboxylase. and metabolic properties. Biosci. Biotechnol. Biochem., Appl. Environ. Microbiol., 61, 4484–4486 (1995). 66, 2474–2478 (2002). 6) Cavin, J.-F., Barthelmebs, L., and Divies, C., Molecular 16) Cavin, J.-F., Barthelmebs, L., Guzzo, J., Van Beeumen, characterization of an inducible p-coumaric acid decar- J., Bart, S., Travers, J. F., and Divies, C., Purification and boxylase from Lactobacillus plantarum: gene cloning, characterization of an inducible p-coumaric acid decar- transcriptional analysis, overexpression in Escherichia boxylase from Lactobacillus plantarum. FEMS Micro- coli, purification, and characterization. Appl. Environ. biol. Lett., 147, 291–295 (1997). Microbiol., 63, 1939–1944 (1997). 17) Huang, Z., Dostal, L., and Rosazza, J. P. N., Purification 7) Cavin, J.-F., Dartois, V., and Divies, C., Gene cloning, and characterization of a ferulic acid decarboxylase from transcriptional analysis, purification, and characteriza- Pseudomonas fluorescens. J. Bacteriol., 176, 5912–5918 tion of phenolic acid decarboxylase from Bacillus (1994). subtilis. Appl. Environ. Microbiol., 64, 1466–1471 18) Zdobnov, E. M., and Apweiler, R., InterProScan: an (1998). integration platform for the signature-recognition meth- 8) Barthelmebs, L., Lecomte, B., Divies, C., and Cavin, ods in InterPro. Bioinformatics, 17, 847–848 (2001). J.-F., Inducible metabolism of phenolic acids in Ped- 19) Barthelmebs, L., Divies, C., and Cavin, J.-F. O., iococcus pentosaceus is encoded by an autoregulated Expression in Escherichia coli of native and chimeric operon which involves a new class of negative tran- phenolic acid decarboxylases with modified enzymatic scriptional regulator. J. Bacteriol., 182, 6724–6731 activities and method for screening recombinant E. coli (2000). strains expressing these enzymes. Appl. Environ. Micro- 9) Overhage, J., Prieffert, H., and Steinbuchel, A., Bio- biol., 67, 1063–1069 (2001). chemical and genetic analyses of ferulic acid catabolism 20) Li, L., Bannantine, J. P., Zhang, Q., Amonsin, A., May, in Pseudomonas sp. strain HR199. Appl. Environ. B. J., Alt, D., Banerji, N., Kanjilal, S., and Kapur, V., Microbiol., 65, 4837–4847 (1999). The complete genome sequence of Mycobacterium 10) Hashidoko, Y., Tanaka, T., and Tahara, S., Induction of avium subspecies paratuberculosis. Proc. Natl. Acad. 4-hydroxycinnamate decarboxylase in Klebsiella oxy- Sci. USA, 102, 12344–12349 (2005). toca cells exposed to substrate and non-substrate 4- 21) Hashidoko, Y., Urashima, M., and Yoshida, T., Predom- hydroxycinnamate analogs. Biosci. Biotechnol. Bio- inant epiphytic bacteria on damaged Polymnia sonchi- chem., 65, 2604–2612 (2001). folia leaves, and their metabolic properties on phenolics 11) Heidelberg, J. F., Eisen, J. A., Nelson, W. C., Clayton, R. of plant origin. Biosci. Biotechnol. Biochem., 58, 1894– Identification of Pad Gene from Klebsiella oxytoca 123 1896 (1994). anomalus. Enzyme Microbiol. Technol., 22, 232–239 22) Clausen, M., Lamb, C. J., Megnet, R., and Doerner, P. (1998). W., PAD1 encodes phenylacrylic acid decarboxylase 24) Rangarajan, E. S., Li, Y., Innuzzi, P., Tocilj, A., Hung, which confers resistance to cinnamic acid in Saccha- L.-W., Matte, A., and Cygler, M., Crystal structure of a romyces cerevisiae. Gene, 142, 107–112 (1994). dodecameric FMN-dependent UbiX-like decarboxylase 23) Edlin, D. A. N., Narbad, A., Gasson, M. J., Dickinson, J. (Pad1) from Escherichia coli O157:H7. Protein Sci., 13, R., and Lloyd, D., Purification and characterization of 3006–3016 (2004). hydroxycinnamate decarboxylase from Brettanomyces